[Cake] [Bloat] dslreports is no longer free
ford at isoc.org
Wed May 27 05:08:41 EDT 2020
What's the bufferbloat verdict on https://speed.cloudflare.com/ ?
> On 1 May 2020, at 20:48, Sebastian Moeller <moeller0 at gmx.de> wrote:
> Hi Dave,
> well, it was a free service and it lasted a long time. I want to raise a toast to Justin and convey my sincere thanks for years of investing into the "good" of the internet.
> Now, the question is which test is going to be the rightful successor?
> Short of running netperf/irtt/iper2/iperf3 on a hosted server, I see lots of potential but none of the tests are really there yet (grievances in now particular order):
> OOKLA: speedtest.net.
> Pros: ubiquitious, allows selection of single flow versus multi-flow test, allows server selection
> Cons: only IPv4, only static unloaded RTT measurement, no control over measurement duration
> BUFFERBLOAT verdict: incomplete, maybe usable as load generator
> NETFLIX: fast.com.
> Pros: allows selection of upload testing, supposedly decent back-end, duration configurable
> allows unloaded, loaded download and loaded upload RTT measurements (but reports sinlge numbers for loaded and unloaded RTT, that are not the max)
> Cons: RTT report as two numbers one for the loaded and one for unloaded RTT, time-course of RTTs missing
> BUFFERBLOAT verdict: incomplete, but oh, so close...
> NPERF: nperf.com
> Pros: allows server selection, RTT measurement and report as time course, also reports average rates and static RTT/jitter for Up- and Download
> Cons: RTT measurement for unloaded only, reported RTT static only , no control over measurement duration
> BUFFERBLOAT verdict: incomplete,
> THINKBROADBAND: www.thinkbroadband.com/speedtest
> Pros: IPv6, reports coarse RTT time courses for all three measurement phases
> Cons: only static unloaded RTT report in final results, time courses only visible immediately after testing, no control over measurement duration
> BUFFERBLOAT verdict: a bit coarse, might work for users within a reasonable distance to the UK for acute de-bloating sessions (history reporting is bad though)
> honorable mentioning:
> BREITBANDMESSUNG: breitbandmessung.de
> Pros: query of contracted internet access speed before measurement, with a scheduler that will only start a test when the backend has sufficient capacity to saturate the user-supplied contracted rates, IPv6 (happy-eyeballs)
> Cons: only static unloaded RTT measurement, no control over measurement duration
> BUFFERBLOAT verdict: unsuitable, exceot as load generator, but the bandwidth reservation feature is quite nice.
> Best Regards
>> On May 1, 2020, at 18:44, Dave Taht <dave.taht at gmail.com> wrote:
>> They ran out of bandwidth.
>> Message to users here:
>> Make Music, Not War
>> Dave Täht
>> CTO, TekLibre, LLC
>> Tel: 1-831-435-0729
>> Cake mailing list
>> Cake at lists.bufferbloat.net
> Bloat mailing list
> Bloat at lists.bufferbloat.net
More information about the Cake